Longitudinal modifications of the MELD score have prognostic meaning in patients with liver cirrhosis.

Background: The Model for End-Stage Liver Disease (MELD) score is an important and well established tool for assessing prognosis in patients with liver cirrhosis. It has been suggested that the longitudinal evaluation of the MELD score may reflect the progression of liver failure more reliably and therefore be more useful in prognostic assessment.

Aim: To assess the prognostic meaning of MELD score modifications in a cohort of cirrhotic patients in whom clinical and biochemical workup was carried out at least twice during a minimum interval of 30 days.

Methods: Forty-six cirrhotic patients were longitudinally evaluated for a median follow-up of 365 days. After initial assessment, all the patients had at least one clinical and biochemical reevaluation during follow-up, which was performed no less than 1 month after initial evaluation. MELD was calculated at entry and at second evaluation. DeltaMELD was calculated as MELD at second evaluation minus MELD at entry. DeltaMELD/time was calculated as DeltaMELD divided by time elapsed between initial assessment and second evaluation expressed in months.

Results: During follow-up, 13 patients died (28%). The median interval between clinical evaluations was 120 days. MELD scores at entry (13 +/- 4 vs 16 +/- 6, P = 0.0516) and DeltaMELD (0 +/- 4 vs 4 +/- 2, P = 0.0028) were significantly different between patients who died and those who survived during the 1-year follow-up. All the patients who died during follow-up showed an increase of at least 1 unit in DeltaMELD/time (sensitivity = 100%), and all the patients who survived showed a decrease of more than 1 unit in DeltaMELD/time (specificity = 100%).

Conclusions: Longitudinal evaluation of the MELD score provides important prognostic information that seems to complete the prognostic definition provided by "static" MELD. Prospective studies in larger series are needed to validate the prognostic use of MELD modifications over time.